Hi, I just read on Distrowatch that the Debian 6.0.2 update has been released so I did aptitude update & then aptitude safe-upgrade & download ~162Mb fine. Before it started to install there was a message that didn't appear important (hence i didn't note it down) as it wasnt an error. I think I spent more time figuring out how to get past it; Enter & Esc didnt work etc, but eventually I found that 'q' got me past it. Anyway it then proceeded to install all the downloaded packages without any problem.

The difficulty I am now experiencing is that the system kernel panic's when booting up. After GRUB it gives the following message:

This info is as noted down with pen & paper as I couldn't find it in any system log.

It doesn't always do this, maybe 50% of the time or so. FYI: my system has been modestly overclocked by only 200Mhz for years now & i have gone back to stock speeds to see if it'd fix the problem, but this didn't.

Any ideas, has anyone else just updated to Debian 6.0.2 & encountered similar problems?

Thanks.

EDIT: I checked the disk via [shutdown -rF now] & it only found 0.5% non-contiguous on the root file system and 3.4% on the file system & it didnt solve the problem.

Squeeze has "apt-listchanges" installed by default. This is what you were getting past.

I suggest installing "apt-listbugs" too. It requires your permission to proceed if there is an outstanding bug.

As for now you will need to use the boot menu for an older kernel and hope that is the problem, not some other package causing the panic.

Also, since it's only happening sporadically, is there anything in the syslog files. Debian usually keeps several old versions of log files before they get overwritten.

Make sure you have the old kernel handy and in the grub menu so you can boot it if you have problems.

Also, are you running mixed sources by any chance? Can you post your /etc/apt/sources.list file? Have you installed any packages from testing or unstable that might be aiding this.

Thanks for the replies, yeah the notification was just something to do with changes, I'll install the apt-listbugs package too, thanks.

Here's my sources.list:

I've not installed any packages from non-squeeze sources - as I've no real need to, plus I don't know how to!

How do I keep the old kernel handy or know if I've got any? I just do the aptitude update & aptitude safe-upgrade to keep my system up to date with security & keeping the kernel up to date. I did try to install the bigmem kernel ages ago becasue I've 4GB of RAM, but I never had any success getting them to work with fglrx so I gave up & just kept on using the bog-standard kernel.

If you are running grub unmodified it will have all of your kernels on it. If Debian is the only OS on your box then the menu is "hidden". You can see it if you hit the space bar when it is supposed to be up. I am not sure what the default time out is (maybe 3 seconds) but you need to be quick on the finger to hit the space bar.

I think you should see your menu every time you boot just for these kinds of situations but on single OS installs hidden is the default. You can fix that by commenting the line relevant to "hidden" in your /etc/default/grub file. You can also change the length of the time out when you are there too if you want.

I set mine at 100 and usually hit enter to make it boot when it has been up about 1 second. I like the option of reading the damned thing (I have several Debian versions installed).

Have you reinstalled your fglrx drivers after the upgrade? Sometimes graphic drivers may glitch if there are major upgrades done to the system.

Hmm, now it doesnt want to do anything other than panic when using the normal kernel or the single-user mode version. I do see the GRUB screen as I dual-boot with windows, so I can do things from there. Thankfully the bigmem kernel is loading up fine, but without fglrx it is quite slow (I've tried reinstalling it on the bigmem kernel but it just blank-screens when trying to load X). Scrolling is horribly slow without fglrx.

I don't think the fglrx drivers are the problem as the system kernel panics long before getting to the stage of loading them; the timestamps on the panic screen are only about 0.07 seconds after GRUB starts the system.

Do you think the kernel itself is at fault, i.e. is the newer one I updated to yesterday via 6.0.2 buggy? Is it possible to go back to the version before this one? When I look in /lib/modules I only see 2.6.32-5-686 & 2.6.32-5-686-bigmem, and in /boot it only mentions files for these two also, I don't think my system has any previous versions stored on it.

EDIT: I've just tried aptitude reinstall linux-image-2.6-686, it didn't download much [4394B only] & it didn't fix the issue.

I've 'aptitude purge linux-image-2.6.32-5-686' and its headers & re-downloaded & installed them again, but still this new non-bigmem Debian 6.0.2 kernel panics when booting with the same screen dumped info as above. I'm able to use the system ok with the bigmem kernel at the moment but it is not much use without fglrx; I tried reinstalling fglrx on it once I'd removed the other kernel image & headers, but it didn't do anything new - it still blanks when trying to load X as per a previous thread of mine http://www.linuxquestions.org/questi...-fglrx-856360/.

I've downloaded the latest Debian 6.0.2 image to see if a fresh install would fix things, but I'd much rather fix this installation than start from scratch again.

Any other thoughts please?

EDIT: I've Googled the 'perf_cpu_notify' but haven't found anything useful from it & also the bit about trying to kill init, but that seems to be a fairly early thing in the boot process apparently. The line 'Pid: 1, comm: swapper Tainted: G D 2.6.32-5-686 #1' isn't getting me anywhere yet so far.

EDIT2: I've just been passing parameters to the non-bigmem kernel via GRUB such as nosplash, nosmp, noacpi and now the system seems to be loading up ok without panicing. Even from cold boot without passing any of those parameters it now loads up... so the problem has currently fixed itself, but I expect it will come back.

Well I have no more thoughts on your problem but I did look at your sig.

Even took a look at the photos. Very nice stuff.

Unfortunately my wife caught me at it. I had to grab several of them for her. I think she likes them too.

Heh, no problem, I hope you enjoy the photos. I've not taken any in quite a while. I hope you mean the arty ones in the sub-folders!

We were most attracted to the monochrome and landscape folders. Some in the landscape stuff could be here in Montana, well not the ones with much water.

My wife really likes the church and the bonnet pictures. I like the rocks myself and the more open country ones.

As a kind of photographer myself I liked the sparkle_grass one a lot. My favorite is the lonely_field_shed though. Something like that can be a real welcome sight if you have been riding (hunting cows probably) for several hours. A place to take a break with some shade.

I know most would see, perhaps, desolation. I see refuge and a promise of a short time of rest for me and my horse.

Thanks for the comments, those ones turned out quite well. All the decent one were taken within Fife, Scotland. I really should get back into photography as a hobby & I've even got some rolls of film still to use up, although they are probably past their use-by date now! I keep thinking about getting a slide scanner, but barely anyone makes them anymore & they cost a lot.

Regarding the linux stuff for anyone still following this, it is still happening & I'll update if I manage to solve it. Yesterday's GRUB meddling with nosplash/nosmp/noacpi doesn't seem to be preventing boot-failure today, but gives a blank screen with cursor rather than the full kernel panic screen.

I've been trying more thing like rebuilding the initramfs, but that didn't work. Anyway I've just removed two 1BG RAM modules, leaving me with 2GB but at least it now prevents the kernel panic/blinking cursor (it would do either depending on how it was feeling).

The total guesswork paragraph: I suspect the updated Squeeze 6.0.2 kernel may handle the memory differently as this problem only arose after the update & my bigmem kernel does not experience the issue at all, maybe it handles memory differently to the standard kernel. I ran the full 4BG through Memtest86 yesterday & no errors were reported. Perhaps I'll see if the two modules I've removed work ok alone in the system, if so it may be that this latest Squeeze 6.0.2 kernel doesn't like having as much as 4GB or RAM in the system.

I've marked the thread as solved for now.

Well that must be a relief anyway.

Strange that it will not handle 3 Gb though.

I drop from 4GB to 2Gb rather than 3GB as that way i don't lose dual-channel mode, and the manual doesnt show 3 modules as a recommended configuration.

EDIT: I've been able to put the memory back in & it works now with 4GB. I initially had two 1GB modules of memory but later added another two 1GB modules to make 4GB. The memory was meant to be exactly the same but upon closer inspection in cpu-z the timings are slightly different although the BIOS was already configured to the lower common setting. Anyway it is working now without kernel panics, the two original & two new modules in being swapped positions.